[SOLVED] rsnapshot - missing last month on a 3-month cycle. Fedora 15

steve_s · 02-16-2012, 04:07 AM

My rsnapshot seems to be running perfectly except that I am a monthly backup too few.
There is no shortage of disk space and I have been running it for long enough to have three monthly backups. I would appreciate any ideas as to why my oldest month is not being retained.

Many thanks in advance.

/etc/crontab contains

Code:

00 */4 * *  * root       /usr/bin/rsnapshot hourly
10 4  *  *  * root       /usr/bin/rsnapshot daily
20 4  *  *  1 root       /usr/bin/rsnapshot weekly
30 4  1  *  * root       /usr/bin/rsnapshot monthly

/etc/rsnapshot.conf contains

Code:

interval        hourly  6
interval        daily   7
interval        weekly  4
interval        monthly 3

destination directory contains

Code:

daily.0  daily.5   hourly.3   weekly.0
daily.1  daily.6   hourly.4   weekly.1
daily.2  hourly.0  hourly.5   weekly.2
daily.3  hourly.1  monthly.0  weekly.3
daily.4  hourly.2  monthly.1

/var/log/rsnapshot contains

Code:

[01/Feb/2012:04:30:01] mv /mnt/volume12/snapshots/monthly.0/ /mnt/volume12/snapshots/monthly.1/
[01/Feb/2012:04:30:01] mv /mnt/volume12/snapshots/weekly.3/ /mnt/volume12/snapshots/monthly.0/

rsnapshot configtest returns

Code:

Syntax OK

rsnapshot -t monthly-t monthly returns

Code:

echo 13561 > /var/run/rsnapshot.pid 
mv /mnt/volume12/snapshots/monthly.1/ /mnt/volume12/snapshots/monthly.2/ 
mv /mnt/volume12/snapshots/monthly.0/ /mnt/volume12/snapshots/monthly.1/ 
mv /mnt/volume12/snapshots/weekly.3/ /mnt/volume12/snapshots/monthly.0/

/var/log/messages contains

Code:

Feb  1 04:30:01 orion rsnapshot[6892]: /usr/bin/rsnapshot monthly: completed successfully

syg00 · 02-16-2012, 09:43 PM

File in use by the first "mv". Second mv gets skipped, ...

steve_s · 02-17-2012, 04:51 AM

I have seen my error: rsnapshot was as comatose as I was at 0430 on January 01. No monthly backup took place, so nothing missing.
I did an out of cycle rsnapshot monthly, and monthly.2 was created nicely.
I have reordered my cron timings to prevent hourly from running into daily, but there had been no collisions anyway.

Code:

00 */4 * *  * root       /usr/bin/rsnapshot hourly
50 3  *  *  * root       /usr/bin/rsnapshot daily
40 3  *  *  1 root       /usr/bin/rsnapshot weekly
30 3  1  *  * root       /usr/bin/rsnapshot monthly

One mystery remains to me however; monthly, weekly, daily and hourly rsnapshot operations all begin and end like,

Code:

mv /mnt/volume12/snapshots/hourly.5/ /mnt/volume12/snapshots/_delete.nnnn/
...
/bin/rm -rf /mnt/volume12/snapshots/_delete.nnnn

and according to /var/log/rsnapshot this has happens every time, except that every single day, 0800 rsnapshot hourly misses out the hourly.5 to _delete.nnnn stage.
I see nothing special about 0800. It seems a it may only cause a minor issue if I ever have to recover from a system crash at 0801.
But what the heck is it about? My original problem is solved (thanks syg00), but if anyone has a thought about this I would welcome it.